Export data

For: analysts, developers, anyone who needs to hand data off.

QueryDeck exports from two places: a query result, and a full table. Both support the same four formats.

Export formats

Format	Best for	Notes
CSV	Spreadsheets, Excel, Google Sheets	Configurable delimiter, quoting, header row
JSON	APIs, JavaScript tools, data pipelines	Array of objects, one per row
SQL	Database-to-database migration	`INSERT` statements with explicit columns
Excel (.xlsx)	Sharing with non-technical teammates	One sheet per result, types preserved

Export from a query result

After running a query, the result grid has an Export button in the top-right of the result pane. Click it.

A sheet opens with format + options:

Format — pick one of CSV, JSON, SQL, Excel.
Rows to export — All rows (re-runs the query without LIMIT), Visible rows (just what's loaded), Selected rows (the ones you've highlighted).
Format-specific options (see below).

Choose a destination, save. The file lands on disk and Finder opens to it.

Export a full table

In Table Mode, click Export in the table toolbar. Same sheet, same options, but All rows means the entire table — filters applied if any are active.

For very large tables (millions of rows), the export streams to disk as it reads from the database. Memory usage stays low. The progress bar tells you how many rows are done.

CSV options

Option	Default	Notes
Delimiter	`,`	Use `;` for European locales, `\t` for TSV
Quote character	`"`	Standard CSV quoting
Quote style	Minimal	Quote only when needed, or always quote
Header row	On	Include column names as the first line
NULL representation	empty	Or a literal string like `NULL` if your downstream tool needs it
Line ending	`\n`	Or `\r\n` for Windows compatibility
Encoding	UTF-8	UTF-8 with BOM available for Excel compatibility

JSON options

Option	Default	Notes
Format	Array of objects	Or one object per line (newline-delimited JSON) for streaming
Indent	Pretty (2 spaces)	Or compact for minimal file size
NULL representation	`null`	Or omit the key entirely
Date format	ISO 8601	Or Unix timestamp

SQL options

Option	Default	Notes
Statement type	`INSERT`	Or `INSERT ... ON CONFLICT DO NOTHING` for idempotent imports
Batch size	100 rows per statement	Larger batches are faster but harder to debug
Table name	The source table	Override to retarget the inserts elsewhere
Include schema	Off	Prefix with `public.` or your schema name

Excel options

Option	Default	Notes
Sheet name	Result name or table name
Number format	General	Or specify a format string per column
Date format	Locale default
Boolean as	TRUE/FALSE	Or 1/0

Excel files preserve column types — booleans become real Excel booleans, dates become real Excel dates.

Copy as

For a single row or selection, you don't need a file. Right-click in the grid → Copy as:

Copy as	What goes on the clipboard
Cell value	The raw value, no formatting
Tab-separated row	Paste straight into a spreadsheet
CSV line	Comma-separated, ready for a CSV file
JSON object	`{ "col1": "value1", "col2": 123 }`
INSERT statement	`INSERT INTO table (col1, col2) VALUES ('value1', 123);`

Cmd+Shift+C is the shortcut for "Copy as INSERT", the most common case during development.

When you've edited rows in Table Mode but haven't committed yet, the pending changes bar has an Export to .sql button. The file contains the exact INSERT, UPDATE, DELETE statements that would run on commit.

Useful for:

Running the same change on staging before committing on production.
Code review with a teammate.
Auditing what an edit session would change.

Limitations

Excel has a hard cap of 1,048,576 rows per sheet. QueryDeck splits across sheets if you exceed it.
SQL exports are not minified or escape-optimized. For very large data dumps, use pg_dump or mysqldump instead.
Binary columns (bytea, BLOB) are exported as base64 in JSON, as escaped strings in CSV, and as decode('...', 'base64') in SQL.

What's next

Inline edits with a diff preview when you want to change rows, not just read them
Notebooks for shareable reports that combine SQL, prose, and result tables